Improvement of the base station layer 1 log analysis with the Rain framework
Tolonen, Tuomas (2022)
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-202204215631
https://urn.fi/URN:NBN:fi:amk-202204215631
Tiivistelmä
The topic of this thesis was to improve base station layer 1 log analysis at Nokia using Rain, which is a company internal log analysis framework. The aim was to reduce the amount of manual log analysis work in software fault investigations so that more time could be allocated for developing new features for the base station software. For the practical work, an automated log analyzer program was developed, which analyzes the content of the log files, and presents the results to the user via different visualization methods.
The key aspects of automated log analysis are log processing and presentation of the results. However, before the processing can be done, the required log files must be selected and prepared. Most often log files are stored in textual format, which means that the processing can be started without bigger preparations. However, the log files covered in this thesis are originally stored in binary format, requiring decoding before they can be processed.
In the log processing phase, it is important to find the most relevant pieces from a large amount of data. Additionally, the structure of the log entries is a significant factor in the log processing. In this thesis, a traditional log analysis method was used, where the most relevant log entries are highlighted according to specific keywords.
The results of the log analysis were presented using the Rain framework, which forms a stable environment both for the fault investigators and automatic log analyzer developers. In addition to the Rain framework, an existing log analyzer – the L1 Analyzer – was utilized as a base for the implementation of the log analyzer program.
The result of the thesis is an automated log analyzer program, which makes the decoded log files easily accessible, and highlights the selected error and warning details from the logs. Additionally, more knowledge was gained on utilizing the Rain framework and about the requirements of base station log
analysis.
The key aspects of automated log analysis are log processing and presentation of the results. However, before the processing can be done, the required log files must be selected and prepared. Most often log files are stored in textual format, which means that the processing can be started without bigger preparations. However, the log files covered in this thesis are originally stored in binary format, requiring decoding before they can be processed.
In the log processing phase, it is important to find the most relevant pieces from a large amount of data. Additionally, the structure of the log entries is a significant factor in the log processing. In this thesis, a traditional log analysis method was used, where the most relevant log entries are highlighted according to specific keywords.
The results of the log analysis were presented using the Rain framework, which forms a stable environment both for the fault investigators and automatic log analyzer developers. In addition to the Rain framework, an existing log analyzer – the L1 Analyzer – was utilized as a base for the implementation of the log analyzer program.
The result of the thesis is an automated log analyzer program, which makes the decoded log files easily accessible, and highlights the selected error and warning details from the logs. Additionally, more knowledge was gained on utilizing the Rain framework and about the requirements of base station log
analysis.